This paper explores the development and implementation of an IoT-based smart home security system enhanced by AI-powered intrusion detection. This system leverages the interconnectedness of IoT devices, such as smart cameras and sensors, to collect real-time data from the home environment. Artificial intelligence algorithms, particularly machine learning and deep learning, analyze this data to identify patterns and anomalies indicative of potential security threats. The system aims to provide homeowners with enhanced security through real-time monitoring, intelligent alerts, and automated responses. We discuss the core concepts, functionalities, challenges, and future trends associated with this technology, emphasizing the importance of addressing privacy and security concerns.
Introduction
The widespread adoption of Internet of Things (IoT) devices in smart homes has introduced both convenience and significant security challenges due to the devices’ limited resources and expanded attack surfaces. Traditional security methods often fall short, prompting researchers to explore machine learning (ML)-based Intrusion Detection Systems (IDS) to detect cyber threats in IoT environments effectively.
This research evaluates various ML algorithms (Naive Bayes, J48, Random Forest, Bagging, K-Star) using the DS2oS dataset, which contains normal and attack data from smart home IoT devices. The study focuses on improving IDS performance by applying feature selection (Information Gain) to reduce dataset dimensionality from 12 to 6 key features, thereby optimizing resource use without sacrificing accuracy.
Experiments showed that removing the timestamp feature (which caused unrealistic overfitting) led to more reliable results. The Random Forest algorithm achieved the highest accuracy (~99.3%) with the reduced 6-feature dataset, demonstrating effective intrusion detection while lowering computational demands — critical for IoT devices. The findings align with existing research, confirming that ML-based IDS can secure IoT smart homes efficiently even with limited data and resources.
Conclusion
In this research, we compared different machine learning (ML) methods to build an intrusion detection system (IDS) for IoT networks, using the DS2oS dataset. We explored the dataset, explained how IDS works, and described the various attack types it contains.We used a feature selection technique called Information Gain to create different versions of the dataset: one with 12 features, one with 6 important features, and one with 11 features (where we removed the timestamp). We then tested these datasets with various ML algorithms.Our results showed that the Random Forest algorithm achieved the best accuracy, reaching 99.42% on the 6-feature dataset. We also found that including the timestamp as a feature led to unrealistic results, so we excluded it.
References
[1] M. Hasan, M. M. Islam, M. I. I. Zarif, and M. Hashem, “Attack and anomaly detection in iot sensors in iot sites using machine learning approaches,” Internet of Things, vol. 7, p. 100059, 2019.
[2] S. Latif, Z. Zou, Z. Idrees, and J. Ahmad, “A novel attack detection scheme for the industrial internet of things using a lightweight random neural network,” IEEE Access, vol. 8, pp. 89 337–89 350, 2020.
[3] P. Kumar, G. P. Gupta, and R. Tripathi, “A distributed ensemble design based intrusion detection system using fog computing to protect the internet of things networks,” Journal of Ambient Intelligence and Humanized Computing, vol. 12, no. 10, pp. 9555–9572, 2021.
[4] M.-O. Pahl and F.-X. Aubet, “All eyes on you: Distributed multi-dimensional iotmicroservice anomaly detection,” in 2018 14th International Conference on Network and Service Management (CNSM). IEEE, 2018, pp. 72–80.
[5] F. Aubet and M. Pahl, “Ds2os traffic traces,” 2018. [Online]. Available: https://www.kaggle.com/datasets/francoisxa/ds2ostraffic
[6] S. Jadhav, H. He, and K. Jenkins, “Information gain directed genetic algorithm wrapper feature selection for credit rating,” Applied Soft Computing, vol. 69, pp. 541–553, 2018.
[7] N. Japkowicz and M. Shah, Evaluating learning algorithms: a classification perspective. Cambridge University Press, 2011.
[8] T. R. Patil, “Mrs. sssherekar,” performance analysis of j48 and j48 classification algorithm for data classification”,” International Journal of Computer Science And Applications, vol. 6, no. 2, 2013.
[9] X. Deng, Q. Liu, Y. Deng, and S. Mahadevan, “An improved method to construct basic probability assignment based on the confusion matrix for classification problem,” Information Sciences, vol. 340, pp. 250–261, 2016.